Agentic benchmark
AA Harvey LAB leaderboard
Artificial Analysis Harvey LAB-AA. Every model the catalog carries a published AA Harvey LAB value for, ranked by that value.
CategoryAgentic
MeasureTask success rate
TasksLegal agent tasks
DifficultyProfessional legal work
An independently evaluated legal-agent benchmark from Artificial Analysis.
AA Harvey LAB ranking
13 models with a published AA Harvey LAB value, ordered by that value, highest first. Models the source has not scored on this benchmark are not listed — they are unmeasured, not last.
| Rank | Model | Provider | Task success rate |
|---|---|---|---|
| 1 |  Kimi K3 | Moonshot AI | 94.6 |
| 2 |  Claude Fable 5 | Anthropic | 93.6 |
| 3 | Claude Opus 5 | Anthropic | 93.5 |
| 4 | Claude Fable 5.1 | Anthropic | 93.0 |
| 5 |  Grok 4.5 | xAI | 92.4 |
| 6 |  Gemini 3.7 Flash | 90.7 | |
| 7 |  MiniMax M3 | MiniMax | 88.4 |
| 8 |  GPT-5.6 Luna | OpenAI | 87.9 |
| 9 | GPT-5.6 Sol | OpenAI | 87.2 |
| 10 | GPT-5.6 Terra | OpenAI | 85.2 |
| 11 |  Nemotron 3 Ultra | NVIDIA | 81.7 |
| 12 |  Mistral Medium 3.5 128B | Mistral | 69.1 |
| 13 | GPT-OSS 120B | OpenAI | 13.9 |
Evidence key: Observed
Rows are ordered by the value Artificial Analysis Harvey LAB-AA Benchmark Leaderboard published, highest first. The source does not state whether a higher value is the better result, so this page does not either: for a benchmark that measures a rate of failure, read the table from the bottom.